Photoelectrochemical Tandem Cells for Solar Water Splittingcourses.washington.edu/mengr599/notes/Journal-Article-3-2-2016.pdf · potentials against the reversible hydrogen electrode,

Photoelectrochemical Tandem Cells for Solar Water SplittingMathieu S. Prevot and Kevin Sivula*

Laboratory for Molecular Engineering of Optoelectronic Nanostructures, Institute of Chemical Sciences and Engineering, EcolePolytechnique Federale de Lausanne, Station 6, 1015 Lausanne, Switzerland

ABSTRACT: In order to be economically competitive with simple “brute force” (i.e., PV+ electrolyzer) strategies or the production of promising solar fuels, like H2, from fossilfuels, a practical photoelectrochemical device must optimize cost, longevity, andperformance. A promising approach that meets these requirements is the combinationof stable and inexpensive oxide semiconductor electrodes in a tandem photo-electrochemical device. In this article, we give an overview of the field including anexamination of the potential solar-to-fuel conversion efficiency expected in a device withrealistic losses. We next discuss recent advances with increasing the performance ofpromising semiconductor electrode materials and highlight how these advances have led tostate-of-the-art solar-to-chemical efficiencies in the 2−3% range in real devices. Challengesfor further optimization are further outlined.

1. INTRODUCTIONThe realization of a sustainable energy economy is contingenton developing an energy vector that can be efficiently(re)generated from renewable sources, easily stored, andstraightforwardly transported to the point of use. Chemicalfuels, which store energy in covalent bonds, are especiallyattractive for this purpose, as the infrastructure for thewidespread use of fossil fuels is already in place. Solar fuelsproduced by photon-to-chemical energy conversion arepromising substances to fulfill this need. In particular,producing H2 from another renewable resource, water, usingonly sunlight as input energy has been thought of as anattractive goal for decades. Indeed, solar hydrogen is consideredthe central element in a multifaceted and interconnectedchemical networka “solar refinery”which will ultimatelyconvert photons from the sun into raw materials for industrialsynthesis and carbon-based chemical fuels.1 The overallelectrochemical reaction of water splitting consists of bothreduction and oxidation half-reactions as follows (with reactionpotentials against the reversible hydrogen electrode, RHE):

Reduction reaction:

+ → =+ − E2H 2e H ( 0.00 V vs RHE)2 red0

Oxidation reaction:

+ → + = −+ + E2H O 4h O 4H ( 1.23 V vs RHE)2 2 ox0

Overall reaction:

→ + Δ = −E2H O 2H O ( 1.23 V)2 2 2

Since the overall reaction has a negative standard potentialenergy must be added for water splitting to occur. We needonly to consider how to use solar energy to supply the energyneed to drive this reaction. While this possibility seemsstraightforward, special attention is needed if this scheme is to

be considered as a viable route for converting solar energy on ascale commensurate with the global energy demand. Forinstance, any system for converting solar energy must find abalance between minimizing system complexity (i.e., devicecost) and maximizing device performance (i.e., energyconversion efficiency and device longevity) given the relativelylow energy density of solar irradiation. Moreover, anyeconomically feasible system for solar hydrogen productionmust also compete with the price of H2 generated fromconventional sources (US$ 2−3 kg−1 for the steam reformingof natural gas).2

Traditionally, the electrical potential required to drive theoverall water splitting reaction can be supplied by any directcurrent (DC) source to an electrolyzer, where the reductionreaction takes place at the cathode and the oxidation reactiontakes place at the anode. While a difference in bias of only 1.23V should be sufficient to split water into H2 and O2, due to theentropic increase needed to drive this process as well as theoverpotential required to overcome kinetics of the oxygen andhydrogen evolution reactions, modern alkaline electrolyzersusually operate at voltages above 1.8 V.3 The most obviousapproach to generate sufficient voltage for water splitting withsolar energy is to connect multiple standard photovoltaic (PV)cells in series. For example, a traditional pn-junction silicon PVproduces a potential of 0.5−0.6 V at its maximum power pointunder standard conditions. Thus, three or four of theseconnected electrically in series would create sufficient potentialto split water (depending on the type of electrolyzeremployed). This “brute force” PV + electrolysis approach islimited by the price and availability of PV devices andelectrolyzers, and currently results in a price of H2 produced

Received: May 29, 2013Revised: July 11, 2013Published: July 26, 2013

Feature Article

pubs.acs.org/JPCC

© 2013 American Chemical Society 17879 dx.doi.org/10.1021/jp405291g | J. Phys. Chem. C 2013, 117, 17879−17893

pubs.acs.org/JPCC

around US$10 kg−1 assuming an installed cost of $2 per peakwatt4 (which is the expected target for installed residential PVby the year 2020 according to the US DOE SunShotinitiative5). While this price can certainly decrease as the PVmarket grows, a major and general drawback comes from thevoltage output of a tradition pn-junction solar cell, which isstrongly dependent on the illumination intensity. This impliesthat the optimum number of cells required to generatesufficient potential is highly variable due to variations in lightintensity caused by haze, cloud cover, or time of day. Thisimportant disadvantage can result inasmuch as a 50% energyloss, and challenges the viability of the PV + electrolysisapproach. Then again, this has also inspired researchers todevelop systems that can potentially avoid this drawback andalso convert solar energy at a price competitive with fossilsources. Developing systems that meet these goals remains amajor challenge, but the severity of the global energy situationhas stimulated the development of many promising routes. Inthis feature, we describe progress toward developing aninexpensive device for the direct conversion of solar energyinto hydrogen fuel using a photoelectrochemical tandem cell.

2. DIRECT SOLAR WATER SPLITTING WITHPHOTOELECTROCHEMICAL CELLS

Compared to the PV + electrolysis approach, a morestraightforward path to light-driven chemical transformationsis to directly interface a semiconductor and the electrolyte as ina photoelectrochemical (PEC) device. Not only does this routeeliminate the need for a separate electrolyzer (reducing systemcomplexity), fewer electron transfer steps are also required,reducing the number of potential loss channels in the system.When a semiconductor is placed in direct contact with anelectrolyte, the equilibration of the semiconductor’s chemicalpotential (the Fermi level) with the oxidation/reductionpotential of the electrolyte can generate an electric (space-charge) field in the semiconductor of width W, at thesemiconductor−liquid junction (SCLJ). Upon the absorptionof a solar photon with energy hν > Eg (the semiconductor bandgap energy), this space-charge field can separate photo-generated electrons in the semiconductor’s conduction bandfrom their corresponding holes in the valence band. In contrastto a traditional photovoltaic cell where the photogeneratedcharge carriers are extracted by electrical contacts into anelectric circuit, excited charge carriers in a PEC electrode travelto the SCLJ to perform redox chemistry directly. For example,with an n-type photoanode under illumination, photogeneratedholes directly oxidize water into O2, while, for a p-typephotocathode, photogenerated electrons can directly reducewater into H2. In either case, the single photoelectrode absorbstwo photons to produce one molecule of H2 (the so-called S2approach), and must generate sufficient external chemicalpotential (free energy, Δμex) with these photons to overcomethe 1.23 V required to split water plus the electrochemicaloverpotentials for the oxidation and reduction reactions, ηoxand ηred, respectively.Historically, interest in performing inexpensive solar water

splitting using a PEC device goes back decades, and was firstdemonstrated using TiO2 as a single semiconductor absorber inan S2 approach over 40 years ago.6 As an oxide semiconductor,TiO2 offers excellent stability in the relatively harsh aqueousconditions of water splitting. However, this material possesses asemiconductor band gap energy, Eg, around 3.2 eV. Thus, itcannot harvest a significant portion of the solar spectrum and

therefore its potential solar-to-hydrogen conversion efficiency isvery low (less than 2%). Because of this, following thedemonstration with TiO2, substantial effort was investedtoward developing ideal PEC materials. To date, the challengeof identifying a single semiconducting material that can offersubstantial solar light harvesting, adequate stability in the harshaqueous electrolytes required, and appropriate conduction andvalence band energy levels straddling the water reduction andoxidation potentials has proven very difficult. However, sincethe water splitting reaction inescapably requires two separatehalf-reactions, it seems obvious to use two light absorbers forPEC water splitting.Indeed, a simple two-photoelectrode approach to construct a

PEC water splitting cell is to use an n-type semiconductorphotoanode together with a p-type semiconductor cathode.The electrochemical energy scheme for this type of device in a“wireless” configuration is shown in Figure 1 for a photoanode/

photocathode type tandem cell with band gap energies Eg1 anEg2, respectively, and assuming Eg1 > Eg2. In this approach withtwo photoelectrodes where each absorbed photon creates oneexcited electron−hole pair, four photons (two in eachabsorber) must be absorbed to create one molecule of H2(called a D4 approach, accordingly). In addition, to optimizesolar photon harvesting, the cells can be placed one on top ofthe other (i.e., in tandem) and photons not absorbed by thefirst cell are transmitted to and absorbed by the second. Tomotivate the development of this type of device, Weber andDignam7 evaluated the potential solar to hydrogen conversionefficiency, ηSTH, of this tandem approach compared to simplyplacing two cells of the same band gap energy side by side.While the evaluation of the side-by-side approach gave an upperlimit of 16.6% with Eg1 = Eg2 = 1.4 eV (a moderate increaseover the 11.6% predicted with a single absorber approach),using the integrated tandem approach, 22% ηSTH was predictedto be possible with Eg1 = 1.8 eV and Eg2 = 1.15 eV. Since it isimportant to know a reasonable value for the maximum

Figure 1. Electron energy scheme of PEC water splitting using a dual-absorber tandem cell. The absorption of a photon (hν) by asemiconductor with a band gap (Eg) creates an electron−hole pair thatcan be separated by the space charge layer to generate aphotopotential (as indicated by the splitting of the Fermi level,broken line). The combination of two absorbers as shown will increasethe free energy (Δμex) available. This free energy must be greater thanthe energy needed for water splitting (1.23 eV) plus the overpotentiallosses at both the anode and the cathode, ηox and ηred, for the watersplitting reaction to occur. Four photons must be absorbed in thismechanism to produce one molecule of H2.

The Journal of Physical Chemistry C Feature Article

dx.doi.org/10.1021/jp405291g | J. Phys. Chem. C 2013, 117, 17879−1789317880

expected solar energy conversion with reasonably foreseenlosses in a real system, a more extensive analysis by Boltonplotted the maximum ηSTH with respect to different absorberband-gaps and realistic losses.8,9 In this work, it was assumedthat each semiconductor absorbs all photons with energy hν >Eg and transmits all photons with energy hν < Eg. No reflectionor scattering losses were included. Following the method ofBolton, if a minimum reasonable value for the energy loss, Uloss= Eg1 + Eg2 − Δμex + ηox + ηred, is chosen to be 2.0 eV (or 1.0eV for each semiconductor), the maximum expected ηSTH as afunction of the chosen band gap energies, Eg1 and Eg2, can bepredicted considering the distribution of photons in standardsolar illumination (AM 1.5G 1000 W m−2). Figure 2a shows a

contour plot of this case based on Bolton’s method using theASTM G173-03 Reference Spectra10 and further assuming thatthe two semiconductors are connected in series (so that thephotocurrent density must be equal).The shapes of the contours (gray lines), which represent

values of Eg1 and Eg2 that result in the same ηSTH, are easilyrationalized. First, no contours are present in the upper left of

the plot, as this is where the set condition Eg1 > Eg2 is notsatisfied. The upper right region, where ηSTH is also not defined,represents semiconductor combinations that do not possesssufficient Δμex for water splitting in a dual absorberconfiguration given the assumed Uloss. The contours also havea shape defined by an abrupt change in direction (e.g., whereEg1 = 2.69 eV and Eg2 = 2.34 eV on the 5% contour) where thedevice limiting the photocurrent switches, indicating that eachabsorber is producing the maximum possible amount ofphotocurrent at the turning point. The boundary defined bythese “current matching” points divides a region to the lowerleft (where the predicted efficiency is limited by the lightabsorption of the top cell) from the upper right (where ηSTH islimited by the bottom cell).Remarkably, even with the large assumed losses in the

tandem cell, a maximum ηSTH of 21.6% is predicted usingoptimum values of Eg1 = 1.89 and Eg2 = 1.34 eV. This is asignificant improvement over the single absorber case, which ispredicted to give only ηSTH of 12.7% under optimum conditions(using Eg = 2.23 eV) with the same assumed losses (1.0 eV perphoton). The photon harvesting of the standard solar spectrausing the optimized tandem cell compared to the singleabsorber approach is shown in Figure 2b. Here, the advantageof the tandem cell can be clearly seen as the top cell harvestsmore of the spectra compared to the single absorber approach.In addition to our revision of Weber and Dignam’s classic

analysis above with large losses, we note that, during thepreparation of this work, Lewis and co-workers publisheddetailed balance calculations that account for the Shockley−Queisser limit on the Δμex of each absorber.11 The maximumSTH efficiency for a water splitting photoanode/photocathodetandem cell was found to be 29.7% with Eg1 = 1.60 and Eg2 =0.95 eV. Thus, while the conditions and losses specified in amodel can lead to a varied maximum predicted ηSTH, it is clearthat a quite attractive solar-to-hydrogen conversion efficiency(at least) over 20% and possibly even up to 30% can beexpected with a tandem PEC cell. However, the optimum bandgap energy range for these maximum efficiencies is narrow, andpresents a significant challenge for materials development.Then again, considering that standard single crystal silicon PVmodules coupled to high-pressure electrolysis give only up toabout 9.3%,12 a PEC tandem cell with over 10% ηSTH is areasonable goal, and this relaxed constraint leaves a widewindow of suitable Eg’s to choose from. Of course, achievingthis still depends on identifying materials with the correct bandgap energies and band positions suitable for the water oxidationand reduction reactions. Notably, this has been accomplishedusing monolithic, epitaxially grown III−V semiconductorsystems and high-performance PEC tandem cells have beendemonstrated with impressive ηSTH up to 12.5%.13 However,the GaInP2 photocathode used quickly corrodes when incontact with the aqueous electrolyte.14 Indeed, realizing apractical device further requires that the materials employed arewidely available, inexpensive to prepare, and stable in a watersplitting PEC cell. This goal has been the focus of our researchat EPFL as well as many other groups.

3. OXIDE SEMICONDUCTORS AS PHOTOANODESWhile many n-type photoanodes have been considered for thewater oxidation half-reaction, identifying a material that remainsstable in the harsh conditions of water oxidation has proved tobe quite difficult. Transition metal oxide semiconductors have,however, excelled in this aspect, with the prototype being

Figure 2. (a) Contour plot (thick gray lines) showing the maximumpredicted ηSTH with AM 1.5G incident radiation (1000 W m−2) and atotal loss, Uloss, set at 2.0 eV as it depends on the chosensemiconductor band gap energies, Eg,i, i = 1, 2 (with Eg1 > Eg2). (b)The benefits of the tandem cell approach are shown through the AM1.5G solar photon flux as a function of wavelength and photon energy.The shaded area of the spectrum represents the photos that could beharvested using a single semiconductor absorber (yellow) and a dualabsorber tandem approach (light brown and purple).



TiO2.6,15 Moreover, solution-based processes like sol−gel, spray

pyrolysis, or hydrothermal routes can be easily used followed bythermal annealing to obtain oxide thin film electrodes withoutexpensive processing techniques. For TiO2, the rutile or anatasephases exhibit a band gap energy in the range 3.0−3.2 eV whichunfortunately limits its potential solar-to-hydrogen conversionefficiency to around 2%, even if used in a current-matchedtandem cell. However, due to its favorable aspects, stability, andactivity for photoelectrochemical water oxidation, an intenseamount of effort has been focused on increasing the lightabsorption of TiO2 with research reports appearing as early as1977.16 However, the lack of success in increasing the energyconversion efficiency of TiO2 using sunlight to above ∼1% overthe last ∼30 years has resulted in skepticism concerning thepotential with this approach.17 Despite this, interest remainsand some recent reports have renewed promise. For example,Chen et al. suggested a potentially successful route toenhancing the light absorption by using hydrogenated anataseTiO2 nanocrystals and tuning the amount of lattice disorders.18

In this way, midgap electronic states were created, accompaniedby a reduced band gap. The resulting black TiO2 nanocrystalsexhibited substantial activity and stability (over 22 days) in thephotocatalytic production of hydrogen from water undersunlight. Another modern strategy to increase light absorptionon TiO2 has been to harness plasmonic resonance.19−21 In onereport, Cronin and co-workers prepared TiO2 photoanodeswith integrated Au nanoparticles and demonstrated a 66-foldincrease in photocurrent using photons with energy 1.96 eV(well below the band gap energy of TiO2).

21 In general,commercializing plasmonically enhanced electrodes becomesmore interesting with the possibility to employ non-noblemetals to reduce electrode cost.22

A perhaps simpler alternative to searching for a method toincrease the light absorption of TiO2 is to employ a differentstable oxide with a smaller band gap energy. Generally, theband gap energy of a transition metal oxide is defined byoxygen 2p states on the valence band and metal d orbitals forthe conduction band. An oxide with a smaller band gap energywill thus most probably have a similar valence band but aconduction band at a lower energy. This is in fact ideal for theuse of transition metal oxides as photoanodes in tandem cells,and while there are numerous binary and hundreds of possibleternary transition metal oxides, relatively few have beenidentified as suitable for the water oxidation reaction. Thesefew notable examples like WO3,

23 BiVO4,24 and Fe2O3

25 havebeen well-developed by a number of groups. The maximumpossible solar-to-hydrogen conversion efficiency as a functionof the oxide band gap energy, assuming it is used as a top celland the same reasonable losses are present that were used forFigure 2, is shown in Figure 3. Typically, research on theseoxide materials has focused on increasing the photocurrentdensity (in units of mA cm−2) measured with respect to anapplied potential (reported versus a known reference) understandard solar illumination. This photocurrent is a good metricof the number of photogenerated charge carriers that reach theSCLJ and participate in an oxidation reaction. As long as theFaradaic efficiency of water oxidation is unity,26 the photo-current corresponds directly to the rate of water splitting. Themaximum photocurrent possible for a photoanode understandard conditions is also shown in Figure 3.As the original low-band gap oxide alternative to titanium

dioxide, tungsten trioxide was first reported shortly after theinitial demonstration with TiO2.

27 With a band gap energy of

2.6−2.8 eV, it could be the top absorber in a tandem cell thatcould potentially convert up to about 6% of the standardsunlight into hydrogen. In the past decade, Augustynski andothers have optimized the solution-based preparation ofnanostructured thin film electrodes of electrodes of WO3 thatachieve incident photon-to-electric current efficiency (IPCE) ofup to 90% and solar photocurrents in the 2−3 mA cm−2 rangewith the application of an external bias (customarily reported atthe water oxidation potential, 1.23 V vs RHE).28 Similar toTiO2, more recent efforts with this material have focused onincreasing its photon harvesting ability. For example, a smallamount of additional photocurrent can also be extracted using aplasmonic approach,29 and doping has been extensivelyinvestigated.30 However, while WO3 remains an importantmodel compound, the high IPCE values already reportedsuggest that the return on further optimization of this materialis limited.Bismuth vanadate is a comparatively more recent material

exploited as a photoanode. The monoclinic scheelite phase ofBiVO4, with a band gap of about 2.4 eV, was shown to bephotoactive for water oxidation in 1999.31 From Figure 3, thismaterial could reach solar-to-hydrogen efficiencies of over 9%in a tandem cell when used with a second absorber of Eg < 1.9eV. Over the past decade, rapid development has shownimprovements of the photocurrent of BiVO4 based photo-anodes resulting mostly from the incorporation of substitu-tional dopants like W32,33 and Mo34 to increase the majoritycarrier transport. Water oxidation photocurrents have beenobserved up to 2.8 mA cm−2 under standard conditions usingMo as a dopant.35 One drawback exhibited by this material hasbeen the large overpotential required to drive the oxygenevolution reaction (OER). To overcome this, OER catalystslike RhO2

35 and cobalt-phosphate (Co-Pi)36−38 have beenapplied to give a remarkable reduction of overpotential andreasonable photocurrent (over 1 mA cm−2 under AM 1.5Gillumination) at potentials under 1.23 V vs RHE. This is shownin Figure 4 with the Co-Pi catalyst where the authors alsocompare the photocurrent with a one-electron sacrificial probe(H2O2

39) to demonstrate that the Co-Pi/BiVO4 electrode

Figure 3. Maximum solar-to-hydrogen conversion efficiency and solarphotocurrent as a function of the top cell band gap energy in a tandemcell with the loss assumptions identical to Figure 2. The band gapenergies and efficiency range for commonly used semiconductor oxidephotoanodes are also shown.



almost completely eliminates losses due to electron−holerecombination at the SCLJ (but recombination in the bulkremains).37 Overall, the relatively low IPCEs observed with thismaterial (<50%) suggest that further research could pushphotocurrents over the 5 mA cm−2 benchmark under standardconditions if the bulk recombination can be reduced.Hematite (α-Fe2O3) is also a promising transition metal

oxide with a band gap of ca. 2.1 eV. This corresponds to amaximum ηSTH of ca. 15% according to Figure 3, if a secondabsorber with Eg2 < 1.5 eV is used. Besides its relatively smallband gap, α-Fe2O3 is the most stable form of iron oxide underambient conditions, and since iron is ubiquitous in the earth’supper crust, hematite is an outstanding candidate for solarenergy conversion on a scale commensurate with the globalenergy demand. Hardee and Bard40 first turned to Fe2O3 as amaterial for water photolysis in 1976 seeking a photoanodematerial that was both stable under anodic polarization andcapable of absorbing light with wavelengths longer than 400nm. Over the next decade, numerous reports were publisheddescribing continuing studies with pure and doped Fe2O3 madeby various routes. Noteworthy photoelectrochemical studieswere performed on Ti doped polycrystalline sintered hematitepellets41,42 and Nb doped single crystals,43 which led to IPCEsup to about 40% at 370 nm and 1.23 V vs RHE in 1 M NaOH.However, the first decade of work identified the manychallenges of employing this material for photoelectrochemicalwater splitting, most notably a large overpotential and slowwater oxidation kinetics,42 a low absorption coefficient,requiring relatively thick films,44,45 poor majority carrierconductivity,43,46 and a short diffusion length of minoritycarriers.41 These limitations suggested that a compact thin filmgeometry was not suitable for hematite, and indeed, modernreports with hematite have shown that photocurrents can bedrastically increased by careful nanostructuring. A review of thiswork has been recently published.25 At EPFL, we havedemonstrated solar photocurrents in excess of 3.0 mA cm−2

at 1.23 V vs RHE with nanostructured films prepared usingatmospheric pressure chemical vapor deposition (APCVD) andcoated with an OER catalyst (IrO2),

47 and comparable

photocurrents have been achieved with a solution-basedcolloidal approach.48,49

Given the poor charge carrier conductivity of Fe2O3, oneapproach that has gained recent success is the extremely thinabsorber (ETA) or host−guest approach. We have firstdemonstrated this using nanostructed WO3.

50 Here, a thinbut conformal layer of the light absorbing iron oxide was coatedon the oxide scaffold, eliminating the requirement of the Fe2O3to transport charges long distances. More recently, photo-currents have been demonstrated over 1 mA cm−2 at 1.23 V vsRHE using atomic layer deposition with this approach on TiSi2“nanonets”,51 and 2.5 mA cm−2 at 1.4 V vs RHE using APCVDfilms on a high surface area conducting Nb:SnO2 scaffold.52

The drastic effect that the scaffold has on the photocurrent inthis approach is shown in Figure 5. However, in general with

this method, the increase in the interfacial area between theiron oxide and the scaffold has been shown to limit thephotocurrent due to recombination.53 This has instigated a lineof research focused on engineering the interface between theiron oxide and the scaffold material.54−56 SiOx,56 SnO2,

53 andGa2O3

55 interfacial layers have been shown as promising buffersthat decrease recombination. Most recently, Hisatomi et al.reported absorbed-photon-to-current efficiency (APCE) ex-ceeding 40% under 400 nm illumination using films of Fe2O3only 9 nm thick. This impressive APCE performance, which

Figure 4. J−V curves measured for a W:BiVO4 photoanode in 0.1 MKPi buffer (pH 8) under 1 sun AM 1.5 front-side illumination (solid)and in the dark (dotted). The red curve describes PEC water oxidationby the W:BiVO4 photoanode, and the green curve describes the PECbehavior after addition of 0.1 M H2O2 to the electrolyte solution. Theblue curve describes PEC water oxidation by the W:BiVO4photoanode after Co-Pi deposition. Reprinted from ref 37.

Figure 5. A “guest” material, hematite, deposited by APCVD into theNb:SnO2 host as evidenced by cross-sectional SEM (a). Current−voltage measurements under AM 1.5 simulated light in 1 M NaOHresulted in high photocurrents for the host−guest structure (black)and negligible photocurrents for the host-only and guest-only controlsamples (gray) (b). Dark currents are shown as dashed lines.Reprinted from ref 52.



rivals the best nanostructure thin films with optimized thicknessover 500 nm, was enabled with a Nb2O5 underlayer depositedby ALD.54 Further development of the ETA approach will nodoubt lead to further increase in photocurrent with Fe2O3based electrodes. For example, further development of ternaryoverlayers like the ZnFe2O4 employed by Choi and co-workersseems particularly promising.57

In addition to the morphology and interfacial recombination,other factors critical to the performance of hematite electrodeshave been shown to be majority carrier transport andrecombination at the SCLJ. The former issue is most typicallyaddressed by substitutional cation doping,58,59 and the latterissue was initially attributed to poor OER kinetics.60 However,most recently, the role of recombination at surface trappingstates has been recognized.61 Surface passivation techniquesusing (non-OER promoting) overlayers like Al2O3

62 andIn2O3

63 have been shown to eliminate some overpotential byreducing Fermi level pinning in these electrodes, and thesubsequent addition of an OER catalyst can further reduce theoverpotential.64 Despite this, a large overpotential remains evenin hematite electrodes with optimized surface treatments, andthis has sparked debate over the nature of the role of the OERcatalyst and the surface passsivating overlayers.60,65 As will alsobe emphasized later on, the further reduction of theoverpotential for water oxidation is of primary importance tohematite for its application in a water splitting device.Advancement on this front will make a large impact on theoverall solar to hydrogen efficiency in real devices. In additionand in general for photoanode research, identifying stable oxidesemiconductors with activity for the water oxidation reactionand a band gap energy around 1.9 eV is also a major goal in thefield.66

4. PHOTOCATHODES ANDPHOTOANODE/PHOTOCATHODE TANDEM CELLS

While numerous n-type semiconductors have been well-studiedas photoanodes for water oxidation, comparatively very littlework has been done on photocathodes for water reduction. Tofunction as a photocathode, a semiconductor has to have aconduction band located cathodic of the water reductionpotential and also exhibit p-type behavior, so, when it isimmersed in water, minority carriers (electrons) are driven tothe surface of the electrode at the SCLJ. In addition, it isnecessary to find materials with a small enough band gapenergy, so a significant portion of the visible light is harvested.Because the reductive side of the cell is not submitted toconditions as harsh as the ones present on the oxidative side(indeed, this electrode is cathodically protected againstoxidation), the search for efficient photocathodes has notbeen limited to oxide materials. The very first photocathode forwater splitting, reported by Yoneyama et al. in 1975, was asingle crystal of p-type GaP, with a band gap of 2.38 eV.67 Sincethis first report, many different semiconductors were suggestedas photocathodes for water reduction. It is possible to groupthese materials in different categories: p-doped phosphides, p-doped chalcogenides, p-doped silicon, and p-doped oxides.Materials from these four categories have been suggested overthe past decades, with increasing performances, in the searchfor low-cost and efficient photocathodes for water reduction.Single-crystalline phosphides, such as p-GaP, p-InP, or p-

GaInP2, are good solar absorbers, with relatively low band gapenergy, making them potential candidates for photoinducedwater reduction. One of the first tandem PEC cells reported,

with an n-SrTiO3 photoanode and a p-GaP photocathodeachieved, an overall 0.67% efficiency.68 In 1982, Heller and co-workers reported that single-crystalline InP (Eg = 1.35 eV),once passivated by an oxide layer and coated with Rh or Re,achieved respectively 13.3 and 11.4% photocathodic conversionefficiencies in 1 M HClO4.

69 Bockris and co-workers laterreported 8.2% overall conversion efficiency for a PEC cell madeof an n-GaAs photoanode and a single-crystalline p-InPphotocathode.70 GaInP2 (Eg = 1.83 eV) has been studied byTurner and co-workers, and was shown to be reasonably stableunder acidic conditions.14 In 1998, they created a device madeof a p−n GaAs photovoltaic junction connected with acrystalline p-GaInP2 photocathode, and achieved a benchamrk12.4% conversion efficiency in 3 M H2SO4.

13 However, despitehigh efficiencies, gallium and indium phosphides are not idealfor inexpensive large-scale production, due to the scarcity oftheir components and the need for them to be single-crystalline.Chalcogenide p-type materials, initially developed for thin

film photovoltaics, have also been explored as potentialphotocathodes for water reduction. Early studies showed thatZnTe (Eg = 2.25 eV) was unstable, whereas single-crystallineCdTe showed a relatively low band gap (Eg = 1.5 eV) and goodstability at cathodic potentials under acidic conditions.71

However, the expensive material achieved relatively lowperformances when used in combination with n-SrTiO3: only0.18% overall conversion efficiency (3.7 times lower than the p-GaP cathode). Finally, Zoski and co-workers reported a Pt-coated tungsten disulfide (p-WS2, Eg = 1.3 eV) electrode with acathodic conversion efficiency of 6−7% under 50 mW cm−2

(632.8 nm) light intensity, in 6 M H2SO4.72 These early studies

have served as an important foundation for the recent increasein interest in chalcogenides for photocathodes. For example,Lewis and co-workers very recently revisited p-WSe2 andcoated a crystalline (Eg = 1.2 eV) photocathode with platinumand ruthenium to achieve cathodic conversion efficiencieshigher than 7%.73 The authors also stressed the importance ofthe purity of the crystalline phase to achieve good minority-carrier diffusion lengths. CuIn1−xGaxSe2 (CIGS) is a modernand interesting chalcogenide because its band gap can be tunedbetween 1.0 and 1.7 eV, by modifying its Ga:In ratio. Domenand co-workers coated a p-CIGS electrode with a thin n-typeCdS layer, and platinized it to produce a 12 mA cm−2 cathodicphotocurrent at pH 9.5, although the light intensity was notspecified.74 The same group later successfully used the sameapproach with p-CuGaSe2 (Eg = 1.65 eV),75 a cheaper materialinitially reported by Miller and co-workers as suitable forphotoinduced hydrogen production,76 with photocurrent ashigh as 13 mA cm−2 in 0.5 M H2SO4 under 1 sun illumination.Finally, a recent report by Matsumura and co-workers showedthat Pt-CdS-coated electrodes of p-CuInS2 (p-CIS, Eg = 1.5 eV)and p-Cu(In,Ga)S2 (p-Ga:CIS, Eg =1.5 eV), deposited by spraypyrolysis, were able to work as photocathodes for waterreduction in 0.1 M europium nitrate.77 Given these recentdemonstrations, it is clear that several chalcogenide materialsare very promising p-type semiconductors for water reduction.Requirements for small band gap and easy-to-process electro-des are met, and it should be possible to reach high efficiencieswith cheap materials, based on abundant elements. While short-term stability studies seem promising with these materials, thelong-term stability of these materials under operation remainsan important issue for further development as well as replacingthe n-type CdS overlayer used to promote charge extraction.



Because of its abundance and small band gap energy (1.12eV), p-Si makes a good candidate for use in a tandem PEC cell.However, we note that, given the losses assumed for theefficiency prediction in Figure 2, a tandem cell will not generatesufficient Δμex with a silicon photocathode unless a photoanodeof Eg > 2.2 eV is used, limiting the conversion efficiency to ca.15% at maximum. Even so, a lot of effort has been placed intoimproving the performances of silicon photocathodes sincetheir first report in 1976.78 An early study from Wrighton’sgroup showed that, when functionalized with a bipyridiniumcompound and platinum acting as catalysts, a p-Si electrodeachieved 5% efficiency when illuminated at 632.8 nm, asopposed to essentially no hydrogen generation for the barematerial. However, the device showed poor stability, with adecrease of 20% in efficiency in the first hour.79 In 1984,Tsubomura and co-workers reported increased efficiency,achieved by diffusing phosphorus into the material from itssurface.80 The resulting n+p-Si material, when coated withplatinum, showed 7.8% conversion efficiency in a 3 M HBrsolution, oxidizing I− from a HI solution at the anode. Again,the photocurrent was reported to decrease quickly, because ofthe formation of reduced species on the surface of theelectrode. More recently, employing nanostructuring andmorphology control techniques have led to important advanceswith Si photocathodes. For example, in 2011, Lewis and co-workers reported a microwire array of platinized n+p-Siprepared by a templated vapor−liquid−solid growth process,which allowed the use of lower-purity Si compared to otherdeposition techniques.81 Increased conversion efficiency of thismaterial compared to pure p-Si was attributed to a higher band

bending at the n+/p interface, causing a higher difference inenergy between the quasi-Fermi levels under illumination.Another approach, developed by Branz and co-workers, reliedon a metal-assisted etching technique to inexpensivelynanostructure the surface of a Si electrode. This was shownto greatly improve the photocathode performance by bothincreasing its surface area and decreasing the amount of lightreflected by the surface (estimated around 25% for a planarsurface).82 Finally, Chorkendorff and co-workers reported agreatly enhanced stability toward oxidation, obtained by coatingthe surface of the n+p-Si electrode with a thin layer of titanium.They also reported that MoSx could be used as an efficientelectrocatalyst with this system.83 Overall, silicon performancesas a photocathode have been recently improved using differentapproaches, morphology control and surface treatments, and todate, it remains one of the most promising materials for a large-scale production due to its very high abundance. Accordingly,researchers have also begun to develop oxide photoanode/p-Siphotocathode tandem cells for overall water splitting. Forexample, very recently, a n-WO3/p-Si device has recently beenreported by Lewis and co-workers, where the semiconductorswere coated on the opposite sides of the same ITO substrate.84

While the photocurrents were low and no conversion efficiencywas reported, this work critically stresses the significance ofsemiconductor interfaces, as these were found to limit theperformance.A last class of materials that have been studied as potential

photocathodes for water reduction is the p-type metal oxides.Although studied very early, it is only recently that theyachieved remarkable conversion efficiencies. Indeed, in their

Figure 6. Top line: (a) Schematic representation of the Cu2O electrode structure. (b) Scanning electron micrograph showing a top view of theelectrode after ALD of ZnO:Al2O3 (20 nm)/TiO2 (11 nm) followed by electrodeposition of Pt nanoparticles. Bottom line: Current−potentialcharacteristics in 1 M Na2SO4 solution, under chopped AM 1.5 light illumination for the bare Cu2O electrode (c) and for the as-deposited Cu2O/ZnO:Al2O3 (20 nm)/11 nm TiO2/Pt electrode (d). E is the electrochemical potential of the electrode. The insets show the respective photocurrenttransients for the electrodes held at 0 V versus RHE in chopped light illumination with N2 purging. Reprinted with permission from ref 92.Copyright 2011 Nature Publishing Group.



early work on a variety of p-type oxides, Hardee and Bardshowed that only p-CuO and p-Bi2O3 produced a cathodicphotocurrent but were also both unstable.85 Benko and co-workers later confirmed that, although having a low band gap(Eg = 1.35 eV) suitable for water reduction, p-CuOdecomposed under illumination.86 The first stable p-typeoxide photocathode was reported by Somorjai and co-workersin 1982. In their work, they used a polycrystalline p-Fe2O3electrode to reduce water, while the oxidation was performed ata polycrystalline n-Fe2O3 electrode. Unfortunately, theperformance of the device was quite poor: only 0.06% overallconversion efficiency in 0.1 M Na2SO4.

87 Another stablematerial, crystalline p-CaFe2O4 (Eg = 1.9 eV), was laterreported by Sato and co-workers but again with relatively poorperformances: only a few μA·cm−2 under 400 mW cm−2

illumination.88 Much more recently, Ishihara and co-workersreported higher photocurrents (200 μA cm−2) for the samematerial used in tandem with an n-TiO2 photoanode.89 In2008, Grimes’ group reported a p−Cu−Ti−O nanotube arraycoupled with an n-TiO2 photoanode.90 The conversionefficiency of the device was reported to be 0.30%. Given thevery few number of stable p-type oxides, one recent strategy hasbeen to afford stability with surface treatments. McFarland andco-workers published the first report of a protected oxidephotocathode in 2003. They used n-TiO2 to protect anelectrodeposited p-type Cu2O (Eg = 2.0 eV) electrode.91 Byusing a thin protective layer of n-TiO2, McFarland’s teamshowed reduced photocorrosion of the electrode withoutdiminution of its performance. The potential of cuprous oxideas a material composed of readily available elements and theability to be deposited at low temperature by scalabletechniques (e.g., electrodeposition) has stimulated the furtherdevelopment of a stabilized Cu2O electrode. Recently,Paracchino et al. reported a highly efficient electrodepositedp-Cu2O electrode, protected by the atomic layer deposition(ALD) of nanolayers of Al-doped ZnO and TiO2, and coatedwith electrodeposited platinum.92,93 The electrode producedphotocurrents as high as 7.6 mA·cm−2 at 0 V vs RHE, understandard illumination, in 1 M Na2SO4 (see Figure 6), and itsstability has been greatly improved, from lasting only secondswithout protecting layers to producing photocurrent for dayswith optimized protection layers. This drastic improvement instability and the high photocurrent it has been shown to delivermakes Cu2O arguably the state-of-the-art p-type oxide, and oneof the most promising low-cost materials reported so far.Additionally, some efforts have been focused on the creation ofprotected p-Cu2O electrodes evolving hydrogen without theneed of expensive cocatalysts. Reisner and co-workerspresented a NiOx-coated Cu2O photocathode, working intandem with an n-WO3 photoanode at pH 6,94 and Wang andco-worker reported a carbon-protected p-Cu2O photocathodewith 0.56% conversion efficiency in 1 M Na2SO4 under AM 1.5illumination.95 Despite its recent success in the literature, itshould be pointed out that, when using Cu2O as the bottomcell in a tandem configuration, the solar-to-hydrogenconversion efficiency is limited to under 9% according toFigure 2. To achieve this performance, the top cell must have aband gap energy of exactly 2.42 eV. If used as a top cell, rather,the situation becomes less restrictive to achieve efficiency over10%, but the availability of photoanodes with suitable band gapenergy between 1.3 and 1.5 eV remains a limitation. Overall,oxide photocathodes, although considered for a long time asunstable and inefficient materials for water reduction, have

recently become increasingly popular with the development ofnanostructuring and protection techniques. Indeed, low-cost,solution-processed oxides represent a very attractive type ofmaterial for a large-scale device production; however, oxideswith high stability and low band gap energy are generally stillneeded.

5. PHOTOANODE/PHOTOVOLTAIC TANDEM CELLSDue to the difficulty in identifying stable p-type cathodes thatcan operate in tandem with an oxide photoelectrode to give ahigh performance PEC device, oxide-photoanode/photovoltaiccell tandem devices have been developed. This approachprovides a good compromise between device performance,complexity, and stability. A photoelectrode/PV tandem devicecan be connected by simply wiring together the photoanodeand PV cell, or prepared in a monolithic geometry, as shown inFigure 7. The operation conditions of a photoanode/PV

tandem device can be predicted by comparing the (two-electrode) current density−voltage characteristics of eachindividual cell under the actual illumination conditions ofeach separate part (given its position in the particular tandemconfiguration).7 Figure 7b shows the current density, J, withrespect to the applied potential (U) for an arbitraryphotovoltaic cell (solid blue curve) and a photoanode (solidred curve) under illumination conditions. The operating

Figure 7. (a) Electron energy scheme of photoelectrode/photovoltaictandem cells with an n-type photoanode coupled to a pn-junctionphotovoltaic cell. (b) The operation conditions of a photoanode/PVtandem device as shown by the idealized current−voltage curves for aPV cell (blue) on top of an n-type photoanode (red) in the dark(broken lines) and under illumination (solid lines). Relevantparameters of the curves are indicated and described in the text.Reprinted with permission from ref 111. Copyright 2010 MaterialsResearch Society.



potential Uop of the tandem cell is defined as the potentialmeasured at the intersection between the PV and the PEC cellcurves. The current density (A m−2) at this point, Jop,represents the electrical current that flows between the twocells and can be used to estimate the overall solar-to-hydrogenconversion efficiency through the following relation:

ηη

=ΔJ G

qE2STHop H

0farad

S

2

where ΔGH2

0 represents the standard free energy of thehydrogen produced (lower heating value, 2.46 eV per moleculeof H2), ηfarad is the Faradaic efficiency for the water splittingreactions, q is the elementary charge (1.6022 × 10−19 C = 1 eVV−1), and ES represents the power flux from the sun incident onthe tandem cell (1000 W m−2 under standard conditions, AM1.5G). The factor of 2 in the denominator represents thestoichiometric relation between one molecule of H2 and thenumber of electrons needed to produce it electrochemicallyfrom H2O. In practice, the J−U properties of the individualcells can be independently varied to maximize Jop and,accordingly, the ηSTH of the device.It should be noted that, in a real device, Jop is not found by

merely maximizing the current output from each cell (denotedas the short circuit current density, Jsc, for the PV cell and theplateau photocurrent density, Jpl, for the photoanode). Uloss andother device nonidealities (such as high series and low shuntresistances) necessitate that the characteristic cell potentials(denoted as the open circuit photopotential, Uoc, and thephotocurrent onset potential, Uon, for the PV and thephotoanode, respectively) also be considered in deviceoptimization. Indeed, these latter parameters become thecritical factors in the operation of a tandem cell. In addition,we note that the careful measurement of the photocurrentdensities under standard conditions is very important for theaccurate prediction of the ηSTH. Standard methods for themeasurement of PEC cells have recently been reported.96

To date, there have been many demonstrations of this typeof device in the literature. For example, a TiO2 photoanode intandem with a thin film PV device based on Cu(In,Ga)Se2/CdSproduced hydrogen at a rate of 0.052 μL s−1 cm−2 duringunassisted solar water splitting (corresponding to an IPCE of1.02%).97 While less interest has been paid toward using theprototypical TiO2 as a photoanode due to its large band gap,this work demonstrated the importance of using optimizedprotective layers (Nb0.03Ti0.97O1.84 in this case) to eliminatecorrosion of the PV cell in the aqueous conditions.Research efforts have recently focused on using promising

transition metal oxides like WO3 and Fe2O3. Miller and co-workers have investigated combining tungsten98 or iron99 oxidephotoanodes in tandem with a-Si:Ge PV. As a drawback, thelarge overpotential for water oxidation (ηox) and the relativelylow Uoc of an a-Si device requires a double junction PV intandem to provide sufficient potential to drive the overall watersplitting reaction (similar to the PV only case where 3 pn-junctions were needed with a-Si). Despite this, 3% ηSTH wasobtained with the WO3/a-Si:Ge/a-Si:Ge device.98 A similardevice for iron oxide with only one PV (Fe2O3/a-Si:Ge) did notsplit water without an external bias, but it was shown that a 0.65V bias “savings” was earned under AM 1.5G illumination.99

Likewise, a 0.3 V reduction in the necessary overpotential wasnoted when Fe2O3 was deposited on n-Si.100

From a practical perspective, the attractive aspects of using awidely available, highly stable, and inexpensively producedphotoanode are diminished when using a tandem componentthat requires relatively expensive processing techniques (e.g.,the plasma-enhanced chemical vapor deposition of a-Si). Thus,many investigations have focused on using next-generationphotovoltaics that can also be fabricated with inexpensive,solution processing methods. The dye-sensitized solar cell(DSSC) is the prototype example101 of this class ofphotovoltaic device and thus has attracted significant attentionfor use in solar water splitting tandem cells with a stablephotoanode.102 One important advantage of using a DSSC overthat of a conventional pn-junction photovoltaic device is thatthe voltage output of the DSSCs is less sensitive with varyingillumination intensity.103 This implies that a photoanode/DSSC tandem cell designed to operate at 1 sun illuminationwill maintain sufficient Δμex for water splitting compared to aphotoanode/pn-junction combination during conditions oflower light intensities (e.g., during hazy or partially cloudydays).The photoanode/DSSC combination was first suggested by

Augustynski and Gratzel104 with WO3 as the photoanodesuggesting that device would be capable of 4.5% ηSTH based onthe plateau photocurrent (Jpl) of WO3.

105 In practice, tandemdevices with WO3 have been constructed by Park and Bard106

and Arakawa et al.107 giving ηSTH’s up to 2.8% at AM 1.5G (100mW cm−2) using Pt as the cathode. However, similar to the a-Sibased devices, two DSSCs connected in series to thephotoanode were necessary to afford overall water splitting.This was accomplished by positioning these two DSSCs side byside behind the WO3 photoanode.This photoanode/2×DSSC architecture does not fundamen-

tally provide a limitation to the possible solar-to-hydrogenconversion efficiency compared to a two-absorber D4 approachfor WO3 or even the more promising Fe2O3, as less than one-third of the available solar photons have a wavelength shorterthan 600 nm and pan-chromatic dyes with high quantumefficiency extending past 900 nm are being developed.108

However, it does present a challenge to device construction asthe two DSCs need to be constructed each with half of theactive area of the photoanode in order to normalize the totalarea of the device.The development of new dyes for the DSSC during the past

decade has initiated work to explore alternative devicearchitectures for optimizing light harvesting in these watersplitting tandem cells. For example, all-organic dyes, such as thesquaraine dye, have a narrow absorption bandwidth extendinginto the far red region of the visible, and have demonstratedsolar power conversion efficiencies over 6% under AM 1.5Gillumination.109,110 Specifically, the dye coded SQ1 absorbslight with wavelengths only between 550 and 700 nm. Inaddition to considerably reducing costs when compared to aruthenium-based dye, Brillet et al. pointed out that two newdistinct tandem cell configurations become accessible whenusing this type of dye.111 The possible configurations are shownin Figure 8. In addition to the standard photoanode/2×DSSCdesign (Figure 8a), a true trilevel device (photoanode/squaraine DSSC/panchromatic DSSC) would also be effective(Figure 8b). This configuration would eliminate the need toconstruct two DSSCs side by side. A second additionalpossibility is to arrange two side-by-side DSSCs in front of ahematite photoanode (Figure 8c). This “front DSSC”configuration is particularly attractive in view of light harvesting



as transition metal oxides (in particular hematite) have highindexes of refraction, which increase reflection. Moreover, thefront DSSC approach eliminates the need to deposit thephotoanode on a transparent conducting glass and aninexpensive metal foil support could instead be used.Brillet et al. determined how these two new photoanode/

DSSC tandem concepts performed compared to the standardarchitecture using nanostructured hematite photoanodes. Thetrilevel tandem architecture (photoanode/squaraine DSSC/panchromatic DSSC) produced the highest operating currentdensity, and thus the highest expected solar-to-hydrogenefficiency of 1.36%. The conventional photoanode/2×DSSCarchitecture gave 1.16%, and the front DSSC configurationresulted in 0.76%. It should be noted that these values were farbelow the expected 3.3% that should have been possible withthe nanostructured hematite photoanodes used. Reduced lightharvesting caused by scattering and reflection were stronglyaffecting the overall conversion efficiency. This is illustrated bythe optical and electronic characteristics of the trilevel tandem

cell presented in Figure 9. Here, it is shown that, while thehematite photoanode only produces photocurrent for λ < 600

nm, the light transmitted to the middle and back cells isseverely attenuated by scattering and reflection losses. Thiscauses the DSSC photocurrent to limit the overall performanceof the tandem cell. While device engineering should improvethe operating current density by reducing reflection, scattering,and resistive losses, DSSCs with higher Uoc’s would enable atrue D4 photoanode/DSSC water splitting tandem cell.Indeed, in 2012, a specifically designed cobalt redox couple

combined with an all-organic dye (coded Y123) gave DSSCswith Uoc > 1.0 V at 1 sun conditions.112 This breakthroughallowed the first demonstration of a D4 photoanode/DSSCwater splitting tandem cell.113 Devices were assembled withboth nanostructured WO3 and Fe2O3 photoandes, and Jop wasmeasured to give ηSTH values of 3.10 and 1.17% with the WO3/DSSC and Fe2O3/DSSC combinations, respectively. An opticalanalysis also compared the predicted photocurrent from the

Figure 8. Layout of three architectures for a tandem cell using ahematite photoanode and two dye-sensitized solar cells in series. (a)“back DSSC” configuration, (b) “trilevel” configuration, and (c) “frontDSSC” configuration. Reprinted with permission from ref 111.Copyright 2010 Materials Research Society.

Figure 9. Optical and electronic characteristics of the “trilevel” tandemcell. Top panel: The solid line shows the incident solar flux. The IPCEof the hematite (dashed line) is convoluted with the solar flux to givethe electron flux in the photoanode (shaded area). Middle panel: Thesolar flux is convoluted with the transmittance of the hematite (dottedline) to give the incident irradiance on the squaraine DSSC (solid boldline). The irradiance is convoluted with the IPCE of the squaraineDSSC (dashed line) to give the electron flux in the intermediatephotovoltaic device (shaded area). Bottom panel: The solar flux isconvoluted with the transmittance of the squaraine dye plus hematite(dotted line) to give the incident irradiance on the black DSSC (solidbold line). The irradiance is convoluted with the IPCE of the blackDSSC (dashed line) to give the electron flux in the bottomphotovoltaic device (shaded area). Reprinted with permission fromref 111. Copyright 2010 Materials Research Society.



integration of IPCE measurements and the actual J−U behaviormeasured in situ in the device. This optical analysis and the J−Ucurves for each device are shown in Figure 10. In the case of theFe2O3/DSSC tandem cell, Jop is far from the plateau region ofthe hematite electrode photocurrent. This results in aperformance far from the maximum obtainable. The limitationof this system is clearly the late onset of the photocurrent in thephotoelectrode, despite the use of state-of-the-art strategies toshift cathodically the onset of the photocurrent by means ofsurface catalysis and passivation. However, in the case of theWO3/DSSC tandem cell, the photocurrent onset is not alimitation and the Jop is very close to the plateau region of thephotoanode (approximately 1000 mV). The limiting factor inthis case is the low photocurrent obtainable by the photoanodedue to the less than ideal absorption capability of tungstentrioxide in the visible region of the solar spectrum. Despite this,the Fe2O3/DSSC device still exhibited a near unity faradaicefficiency and a good stability over an 8 h testing period.Overall, this work suggests that the low ηSTH of the D4hematite/DSSC tandem cell offers the larger room forimprovement, in particular, further reduction of the over-potential for water oxidation.

6. CONCLUSIONS AND OUTLOOK

Since the discovery of water photoelectrolysis as an attractiveway to produce solar hydrogen, numerous cell designs havebeen suggested and investigated. Due to limitations inidentifying ideal materials, the original single-absorber (S2)approach to achieve direct solar energy conversion has givenway to the more efficient dual-absorber (D4) configuration.Using two absorbing materials indeed allows for better lightharvesting properties, and better stability as the two sides of thecell are subject to different conditions and can be optimizedseparately. It has been further shown that placing twophotoelectrodes one on top of the other allowed for highertheoretical maximum efficiencies than a “side-by-side” design

(21.6 vs 16.6%), notably thanks to a reduction by half of thesurface area of the device.The practical realization of a wireless integrated tandem

device composed of two semiconductor photoelectrodes comeswith different issues to overcome, among which are the stabilityof the photoelectrodes, the poor kinetics of the water splittinghalf-reactions, and the cost of production. However, recentadvances in protection and nanostructuring techniques haveallowed the creation of increasingly efficient and robustphotoelectrodes based on earth-abundant materials, through abetter control of the size and shape of the materials on amicroscopic level. However, long-term stability and the furtherreduction of electrochemical overpotentials remain majorchallenges in the development of photoelectrodes. Thedevelopment of photocathode materials is comparablyimmature to that of their photoanode counterparts, and theidentification of intrinsically stable photocathode material witha band gap energy in the 1.2−1.4 eV range is urgently needed.An alternative tandem-cell design is to combine a photo-

electrode with a photovoltaic device. This approach was initiallynot favored, due to the need to have several photovoltaicjunctions to achieve a high enough voltage to split water, theoften poor stability of the materials in aqueous conditions, andtheir relatively high cost. However, recent progress with high-voltage dye-sensitized solar cells has allowed the demonstrationof a cheap photo-electrocatalytic cell, composed of one oxidephotoanode and one DSSC connected together.After 40 years of research, and thanks to great advances in

the fields of nanotechnology, materials science, and electro-chemistry, the goal of an efficient and stable tandem cell basedon cheap and abundant materials seems to be within reach.While long-term stability and higher solar-to-hydrogenconversion efficiency are still needed, the recent and rapiddevelopment of new photoelectrode materials like BiVO4, androbust, strongly absorbing organic dyes for the DSSC,114 alongwith the continued understanding of the photoelectrochemicalprocesses occurring at the electrodes,60 suggest that these

Figure 10. General scheme (left) of a photoanode/dye-sensitized-solar-cell tandem cell. Spectral response and J−V characteristics of the WO3(yellow)/DSSC (blue) (a, c) and Fe2O3 (red)/DSSC (blue) (b, d) tandem cells. The transmittance of the photoanode (not shown) convoluted tothe AM 1.5G photon flux on the photoanode (a, b black lines) allows the calculation of the photon flux incident to the DSSC (a, b gray lines). IPCEdata (not shown) and the photon flux incident to each element are used to estimate the photocurrent density (shaded areas under the curves in partsa and b). J−V curves (c and d) of the cells are shown under AM 1.5G irradiation. The filled lines represent the J−V curves predicted from calculation(see ref), and the triangles represent data from in situ device measurements. Adapted from ref 113.



remaining challenges can be overcome. The next decades willlikely bring PEC development to economically viable solar fuelproduction by water splitting and the commercialization of thistechnology on which global-scale solar energy storage and“solar refineries” can be based.

■ AUTHOR INFORMATIONCorresponding Author*E-mail: [email protected]. Phone: +41 21 693 79 79.NotesThe authors declare no competing financial interest.Biographies

Mathieu S. Prevot obtained his MSc in matter sciences from the EcoleNormale Superieure de Lyon in 2012 after two research internships atthe University of Calgary. He was accepted as a PhD student in theLaboratory for Molecular Engineering of Optoelectronic Nanomateri-als at EPFL in 2013.

Kevin Sivula obtained a PhD in chemical engineering from UCBerkeley in 2007. In 2012, after leading a research group in theLaboratory of Photonics and Interfaces at EPFL, he was appointedtenure track assistant professor. He now heads the Laboratory forMolecular Engineering of Optoelectronic Nanomaterials (http://limno.epfl.ch) at EPFL.

■ ACKNOWLEDGMENTSThe Faculty of Basic Sciences at EPFL and the EOS holdings(contract number 536067) are acknowledged for providingfinancial support.

■ REFERENCES(1) Schlogl, R. The Role of Chemistry in the Energy Challenge.ChemSusChem 2010, 3, 209−222.

(2) Olateju, B.; Kumar, A. Techno-Economic Assessment ofHydrogen Production from Underground Coal Gasification (UCG)in Western Canada with Carbon Capture and Sequestration (CCS) forUpgrading Bitumen from Oil Sands. Appl. Energy 2013, 111, 428−440.(3) Dutta, S. Technology Assessment of Advanced ElectrolyticHydrogen Production. Int. J. Hydrogen Energy 1990, 15, 379−386.(4) Committee on Alternatives and Strategies for Future HydrogenProduction and Use, N. R. C., National Academy of Engineering. TheHydrogen Economy: Opportunities, Costs, Barriers, and R&D Needs;National Academies Press: Washington, DC, 2004.(5) Goodrich, A.; James, T.; Woodhouse, M. Residential,Commercial, and Utility-Scale Photovoltaic (PV) System Prices inthe United States: Current Drivers and Cost Reduction Opportunities;NREL/TP-6A20-53347L; National Renewable Energy Laboratory:Golden, CO, 2012.(6) Fujishima, A.; Honda, K. Electrochemical Photolysis of Water at aSemiconductor Electrode. Nature 1972, 238, 37−39.(7) Weber, M. F.; Dignam, M. J. Efficiency of Splitting Water withSemiconducting Photoelectrodes. J. Electrochem. Soc. 1984, 131,1258−1265.(8) Bolton, J. R.; Strickler, S. J.; Connolly, J. S. Limiting andRealizable Efficiencies of Solar Photolysis of Water. Nature 1985, 316,495−500.(9) Bolton, J. R. Solar Photoproduction of Hydrogen: A Review.Solar Energy 1996, 57, 37−50.(10) ASTM Standard G173 - 03, 2008, “Standard Tables forReference Solar Spectral Irradiances: Direct Normal and Hemi-spherical on 37° Tilted Surface”, ASTM International, WestConshohocken, PA, 2008, DOI: 10.1520/G0173-03R08, www.astm.org.(11) Hu, S.; Xiang, C.; Haussener, S.; Berger, A. D.; Lewis, N. S. AnAnalysis of the Optimal Band Gaps of Light Absorbers in IntegratedTandem Photoelectrochemical Water-Splitting Systems. EnergyEnviron. Sci. 2013, DOI: 10.1039/c3ee40453f.(12) Kelly, N.; Gibson, T.; Ouwerkerk, D. A Solar-Powered, High-Efficiency Hydrogen Fueling System Using High-Pressure Electrolysisof Water: Design and Initial Results. Int. J. Hydrogen Energy 2008, 33,2747−2764.(13) Khaselev, O.; Turner, J. A. A Monolithic Photovoltaic-Photoelectrochemical Device for Hydrogen Production via WaterSplitting. Science 1998, 280, 425−427.(14) Khaselev, O.; Turner, J. A. Electrochemical Stability of p-GaInP2

in Aqueous Electrolytes toward Photoelectrochemical Water Splitting.J. Electrochem. Soc. 1998, 145, 3335−3339.(15) Fujishima, A.; Zhang, X. T.; Tryk, D. A. TiO2 Photocatalysis andRelated Surface Phenomena. Surf. Sci. Rep. 2008, 63, 515−582.(16) Ghosh, A. K.; Maruska, H. P. Photoelectrolysis of Water inSunlight with Sensitized Semiconductor Electrodes. J. Electrochem. Soc.1977, 124, 1516−1522.(17) Nowotny, J.; Bak, T.; Nowotny, M. K.; Sheppard, L. R.Titanium Dioxide for Solar-Hydrogen I. Functional Properties. Int. J.Hydrogen Energy 2007, 32, 2609−2629.(18) Chen, X.; Liu, L.; Yu, P. Y.; Mao, S. S. Increasing SolarAbsorption for Photocatalysis with Black Hydrogenated TitaniumDioxide Nanocrystals. Science 2011, 331, 746−750.(19) Zhang, Q.; Lima, D. Q.; Lee, I.; Zaera, F.; Chi, M.; Yin, Y. AHighly Active Titanium Dioxide Based Visible-Light Photocatalystwith Nonmetal Doping and Plasmonic Metal Decoration. Angew.Chem., Int. Ed. 2011, 50, 7088−7092.(20) Linic, S.; Christopher, P.; Ingram, D. B. Plasmonic-MetalNanostructures for Efficient Conversion of Solar to Chemical Energy.Nat. Mater. 2011, 10, 911−921.(21) Liu, Z.; Hou, W.; Pavaskar, P.; Aykol, M.; Cronin, S. B. PlasmonResonant Enhancement of Photocatalytic Water Splitting UnderVisible Illumination. Nano Lett. 2011, 11, 1111−1116.(22) Zhou, X.; Liu, G.; Yu, J.; Fan, W. Surface Plasmon Resonance-Mediated Photocatalysis by Noble Metal-Based Composites underVisible Light. J. Mater. Chem. 2012, 22, 21337−21354.



mailto:[email protected]

http://limno.epfl.ch

http://limno.epfl.ch

www.astm.org

www.astm.org

(23) Sartoretti, C. J.; Alexander, B. D.; Solarska, R.; Rutkowska, W.A.; Augustynski, J.; Cerny, R. Photoelectrochemical Oxidation ofWater at Transparent Ferric Oxide Film Electrodes. J. Phys. Chem. B2005, 109, 13685−13692.(24) Sayama, K.; Nomura, A.; Zou, Z.; Abe, R.; Abe, Y.; Arakawa, H.Photoelectrochemical Decomposition of Water on NanocrystallineBiVO4 Film Electrodes under Visible Light. Chem. Commun.(Cambridge, U. K.) 2003, 0, 2908−2909.(25) Sivula, K.; Le Formal, F.; Gratzel, M. Solar Water Splitting:Progress Using Hematite (α-Fe2O3) Photoelectrodes. ChemSusChem2011, 4, 432−449.(26) Mi, Q.; Zhanaidarova, A.; Brunschwig, B. S.; Gray, H. B.; Lewis,N. S. A Quantitative Assessment of the Competition between Waterand Anion Oxidation at WO3 Photoanodes in Acidic AqueousElectrolytes. Energy Environ. Sci. 2012, 5, 5694−5700.(27) Butler, M. A. Photoelectrolysis and Physical-Properties ofSemiconducting Electrode WO3. J. Appl. Phys. 1977, 48, 1914−1920.(28) Solarska, R.; Alexander, B. D.; Braun, A.; Jurczakowski, R.;Fortunato, G.; Stiefel, M.; Graule, T.; Augustynski, J. Tailoring theMorphology of WO3 Films with Substitutional Cation Doping: Effecton the Photoelectrochemical Properties. Electrochim. Acta 2010, 55,7780−7787.(29) Solarska, R.; Krolikowska, A.; Augustyn ski, J. Silver NanoparticleInduced Photocurrent Enhancement at WO3 Photoanodes. Angew.Chem., Int. Ed. 2010, 49, 7980−7983.(30) Janaky, C.; Rajeshwar, K.; de Tacconi, N. R.; Chanmanee, W.;Huda, M. N. Tungsten-Based Oxide Semiconductors for SolarHydrogen Generation. Catal. Today 2013, 199, 53−64.(31) Kudo, A.; Omori, K.; Kato, H. A Novel Aqueous Process forPreparation of Crystal Form-Controlled and Highly Crystalline BiVO4

Powder from Layered Vanadates at Room Temperature and ItsPhotocatalytic and Photophysical Properties. J. Am. Chem. Soc. 1999,121, 11459−11467.(32) Ye, H.; Lee, J.; Jang, J. S.; Bard, A. J. Rapid Screening of BiVO4-Based Photocatalysts by Scanning Electrochemical Microscopy(SECM) and Studies of Their Photoelectrochemical Properties. J.Phys. Chem. C 2010, 114, 13322−13328.(33) Li, M.; Zhao, L.; Guo, L. Preparation and PhotoelectrochemicalStudy of BiVO4 Thin Films Deposited by Ultrasonic Spray Pyrolysis.Int. J. Hydrogen Energy 2010, 35, 7127−7133.(34) Yao, W.; Iwai, H.; Ye, J. Effects of Molybdenum Substitution onthe Photocatalytic Behavior of BiVO4. Dalton Trans. 2008, 1426−1430.(35) Luo, W.; Yang, Z.; Li, Z.; Zhang, J.; Liu, J.; Zhao, Z.; Wang, Z.;Yan, S.; Yu, T.; Zou, Z. Solar Hydrogen Generation from Seawaterwith a Modified BiVO4 Photoanode. Energy Environ. Sci. 2011, 4,4046−4051.(36) Pilli, S. K.; Furtak, T. E.; Brown, L. D.; Deutsch, T. G.; Turner,J. A.; Herring, A. M. Cobalt-Phosphate (Co-Pi) Catalyst Modified Mo-Doped BiVO4 Photoelectrodes for Solar Water Oxidation. EnergyEnviron. Sci. 2011, 4, 5028−5034.(37) Zhong, D. K.; Choi, S.; Gamelin, D. R. Near-CompleteSuppression of Surface Recombination in Solar Photoelectrolysis by“Co-Pi” Catalyst-Modified W:BiVO4. J. Am. Chem. Soc. 2011, 133,18370−18377.(38) Abdi, F. F.; van de Krol, R. Nature and Light Dependence ofBulk Recombination in Co-Pi-Catalyzed BiVO4 Photoanodes. J. Phys.Chem. C 2012, 116, 9398−9404.(39) Dotan, H.; Sivula, K.; Gratzel, M.; Rothschild, A.; Warren, S. C.Probing the Photoelectrochemical Properties of Hematite (α-Fe2O3)Electrodes Using Hydrogen Peroxide as a Hole Scavenger. EnergyEnviron. Sci. 2011, 4, 958−964.(40) Hardee, K. L.; Bard, A. J. Semiconductor Electrodes 5.Application of Chemically Vapor-Deposited Iron-Oxide Films toPhotosensitized Electrolysis. J. Electrochem. Soc. 1976, 123, 1024−1026.(41) Kennedy, J. H.; Frese, K. W. Photo-Oxidation of Water at alpha-Fe2O3 Electrodes. J. Electrochem. Soc. 1978, 125, 709−714.

(42) Dareedwards, M. P.; Goodenough, J. B.; Hamnett, A.;Trevellick, P. R. Electrochemistry and Photoelectrochemistry ofIron(III) Oxide. J. Chem. Soc., Faraday Trans. 1983, 79, 2027−2041.(43) Sanchez, C.; Sieber, K. D.; Somorjai, G. A. The Photo-electrochemistry of Niobium Doped Alpha-Fe203. J. Electroanal. Chem.1988, 252, 269−290.(44) Itoh, K.; Bockris, J. O. Stacked Thin-Film Photoelectrode UsingIron-Oxide. J. Appl. Phys. 1984, 56, 874−876.(45) Itoh, K.; Bockris, J. O. Thin-Film Photoelectrochemistry - Iron-Oxide. J. Electrochem. Soc. 1984, 131, 1266−1271.(46) Horowitz, G. Capacitance Voltage Measurements and Flat-BandPotential Determination on Zr-Doped Alpha-Fe2O3 Single-CrystalElectrodes. J. Electroanal. Chem. 1983, 159, 421−436.(47) Tilley, S.; Cornuz, M.; Sivula, K.; Gratzel, M. Light-InducedWater Splitting with Hematite: Improved Nanostructure and IridiumOxide Catalysis. Angew. Chem., Int. Ed. 2010, 49, 6405−6408.(48) Brillet, J.; Gratzel, M.; Sivula, K. Decoupling Feature Size andFunctionality in Solution-Processed, Porous Hematite Electrodes forSolar Water Splitting. Nano Lett. 2010, 10, 4155−4160.(49) Goncalves, R. H.; Lima, B. H. R.; Leite, E. R. MagnetiteColloidal Nanocrystals: A Facile Pathway to Prepare MesoporousHematite Thin Films for Photoelectrochemical Water Splitting. J. Am.Chem. Soc. 2011, 133, 6012−6019.(50) Sivula, K.; Le Formal, F.; Gratzel, M. WO3-Fe2O3 Photoanodesfor Water Splitting: A Host Scaffold, Guest Absorber Approach. Chem.Mater. 2009, 21, 2862−2867.(51) Lin, Y.; Zhou, S.; Sheehan, S. W.; Wang, D. Nanonet-BasedHematite Heteronanostructures for Efficient Solar Water Splitting. J.Am. Chem. Soc. 2011, 133, 2398−2401.(52) Stefik, M.; Cornuz, M.; Mathews, N.; Hisatomi, T.; Mhaisalkar,S.; Gratzel, M. Transparent, Conducting Nb:SnO2 for Host−GuestPhotoelectrochemistry. Nano Lett. 2012, 12, 5431−5435.(53) Liang, Y. Q.; Enache, C. S.; van de Krol, R. Photo-electrochemical Characterization of Sprayed Alpha-Fe2O3 ThinFilms: Influence of Si Doping and SnO2 Interfacial Layer. Int. J.Photoenergy 2008, 739864.(54) Hisatomi, T.; Dotan, H.; Stefik, M.; Sivula, K.; Rothschild, A.;Gratzel, M.; Mathews, N. Enhancement in the Performance ofUltrathin Hematite Photoanode for Water Splitting by an OxideUnderlayer. Adv. Mater. 2012, 24, 2699−2702.(55) Hisatomi, T.; Brillet, J.; Cornuz, M.; Le Formal, F.; Tetreault,N.; Sivula, K.; Gratzel, M. A Ga2O3 Underlayer as an IsomorphicTemplate for Ultrathin Hematite Films toward Efficient Photo-electrochemical Water Splitting. Faraday Discuss. 2012, 155, 223−232.(56) Le Formal, F.; Gratzel, M.; Sivula, K. Controlling Photoactivityin Ultrathin Hematite Films for Solar Water-Splitting. Adv. Funct.Mater. 2010, 20, 1099−1107.(57) McDonald, K. J.; Choi, K.-S. Synthesis and Photoelectrochem-ical Properties of Fe2O3/ZnFe2O4 Composite Photoanodes for Usein Solar Water Oxidation. Chem. Mater. 2011, 23, 4863−4869.(58) Sivula, K.; Zboril, R.; Le Formal, F.; Robert, R.; Weidenkaff, A.;Tucek, J.; Frydrych, J.; Gratzel, M. Photoelectrochemical WaterSplitting with Mesoporous Hematite Prepared by a Solution-BasedColloidal Approach. J. Am. Chem. Soc. 2010, 132, 7436−7444.(59) Frydrych, J.; Machala, L.; Tucek, J.; Siskova, K.; Filip, J.;Pechousek, J.; Safarova, K.; Vondracek, M.; Seo, J. H.; Schneeweiss,O.; Gratzel, M.; Sivula, K.; Zboril, R. Facile Fabrication of Tin-DopedHematite Photoelectrodes - Effect of Doping On Magnetic Propertiesand Performance for Light-Induced Water Splitting. J. Mater. Chem.2012, 22, 23232−23239.(60) Sivula, K. Metal Oxide Photoelectrodes for Solar FuelProduction, Surface Traps, and Catalysis. J. Phys. Chem. Lett. 2013,4, 1624−1633.(61) Klahr, B. M.; Hamann, T. W. Current and Voltage LimitingProcesses in Thin Film Hematite Electrodes. J. Phys. Chem. C 2011,115, 8393−8399.(62) Le Formal, F.; Tetreault, N.; Cornuz, M.; Moehl, T.; Gratzel,M.; Sivula, K. Passivating Surface States on Water Splitting HematitePhotoanodes with Alumina Overlayers. Chem. Sci. 2011, 2, 737−743.



(63) Hisatomi, T.; Le Formal, F.; Cornuz, M.; Brillet, J.; Tetreault,N.; Sivula, K.; Gratzel, M. Cathodic Shift in Onset Potential of SolarOxygen Evolution on Hematite by 13-Group Oxide Overlayers. EnergyEnviron. Sci. 2011, 4, 2512−2515.(64) Le Formal, F.; Sivula, K.; Gratzel, M. The TransientPhotocurrent and Photovoltage Behavior of a Hematite Photoanodeunder Working Conditions and the Influence of Surface Treatments. J.Phys. Chem. C 2012, 116, 26707−26720.(65) Gamelin, D. R. Water Splitting: Catalyst or Spectator? Nat.Chem. 2012, 4, 965−967.(66) Osterloh, F. E.; Parkinson, B. A. Recent Developments in SolarWater-Splitting Photocatalysis. MRS Bull. 2011, 36, 17−22.(67) Yoneyama, H.; Sakamoto, H.; Tamura, H. A Photo-Electochemical Cell with Production of Hydrogen and Oxygen by aCell Reaction. Electrochim. Acta 1975, 20, 341−345.(68) Ohashi, K.; McCann, J.; Bockris, J. O. M. Stable Photo-electrochemical Cells for the Splitting of Water. Nature 1977, 266,610−611.(69) Aharon-Shalom, E.; Heller, A. Efficient p-InP(Rh-Halloy) and p-InP(Re-Halloy) Hydrogen Evolving Photocathodes. J. Electrochem. Soc.1982, 129, 2865−2866.(70) Kainthla, R. C.; Zelenay, B.; Bockris, J. O. M. SignificantEfficiency Increase in Self-Driven Photoelectrochemical Cell for WaterPhotoelectrolysis. J. Electrochem. Soc. 1987, 134, 841−845.(71) Ohashi, K.; Uosaki, K.; Bockris, J. O. M. Cathodes forPhotodriven Hydrogen Generators: Znte and Cdte. Int. J. Energy Res.1977, 1, 25−30.(72) Baglio, J. A.; Calabrese, G. S.; Harrison, D. J.; Kamieniecki, E.;Ricco, A. J.; Wrighton, M. S.; Zoski, G. D. ElectrochemicalCharacterization of p-Type Semiconducting Tungsten DisulfidePhotocathodes: Efficient Photoreduction Processes at Semiconduc-tor/Liquid Electrolyte Interfaces. J. Am. Chem. Soc. 1983, 105, 2246−2256.(73) McKone, J. R.; Pieterick, A. P.; Gray, H. B.; Lewis, N. S.Hydrogen Evolution from Pt/Ru-Coated p-Type WSe2 Photo-cathodes. J. Am. Chem. Soc. 2013, 135, 223−231.(74) Yokoyama, D.; Minegishi, T.; Maeda, K.; Katayama, M.; Kubota,J.; Yamada, A.; Konagai, M.; Domen, K. Photoelectrochemical WaterSplitting Using a Cu(In,Ga)Se2 Thin Film. Electrochem. Commun.2010, 12, 851−853.(75) Moriya, M.; Minegishi, T.; Kumagai, H.; Katayama, M.; Kubota,J.; Domen, K. Stable Hydrogen Evolution from CdS-ModifiedCuGaSe2 Photoelectrode under Visible-Light Irradiation. J. Am.Chem. Soc. 2013, 135, 3733−3735.(76) Marsen, B.; Cole, B.; Miller, E. L. Photoelectrolysis of WaterUsing Thin Copper Gallium Diselenide Electrodes. Sol. Energy Mater.Sol. Cells 2008, 92, 1054−1058.(77) Ikeda, S.; Nonogaki, M.; Septina, W.; Gunawan, G.; Harada, T.;Matsumura, M. Fabrication of CuInS2 and Cu(In,Ga)S2 Thin Films bya Facile Spray Pyrolysis and Their Photovoltaic and Photo-electrochemical Properties. Catal. Sci. Technol. 2013, 3, 1849−1854.(78) Candea, R. M.; Kastner, M.; Goodman, R.; Hickok, N.Photoelectrolysis of Water: Si in Salt Water. J. Appl. Phys. 1976, 47,2724−2726.(79) Bookbinder, D. C.; Bruce, J. A.; Dominey, R. N.; Lewis, N. S.;Wrighton, M. S. Synthesis and Characterization of a PhotosensitiveInterface for Hydrogen Generation: Chemically Modified P-TypeSemiconducting Silicon Photocathodes. Proc. Natl. Acad. Sci. U.S.A.1980, 77, 6280−6284.(80) Nakato, Y.; Egi, Y.; Hiramoto, M.; Tsubomura, H. HydrogenEvolution and Iodine Reduction on an Illuminated n-p JunctionSilicon Electrode and Its Application to Efficient Solar Photo-electrolysis of Hydrogen Iodide. J. Phys. Chem. 1984, 88, 4218−4222.(81) Boettcher, S. W.; Warren, E. L.; Putnam, M. C.; Santori, E. A.;Turner-Evans, D.; Kelzenberg, M. D.; Walter, M. G.; McKone, J. R.;Brunschwig, B. S.; Atwater, H. A.; Lewis, N. S. PhotoelectrochemicalHydrogen Evolution Using Si Microwire Arrays. J. Am. Chem. Soc.2011, 133, 1216−1219.

(82) Oh, J.; Deutsch, T. G.; Yuan, H.-C.; Branz, H. M. NanoporousBlack Silicon Photocathode for H2 Production by Photoelectrochem-ical Water Splitting. Energy Environ. Sci. 2011, 4, 1690−1694.(83) Seger, B.; Laursen, A. B.; Vesborg, P. C. K.; Pedersen, T.;Hansen, O.; Dahl, S.; Chorkendorff, I. Hydrogen Production Using aMolybdenum Sulfide Catalyst on a Titanium-Protected n+p-SiliconPhotocathode. Angew. Chem., Int. Ed. 2012, 51, 9128−9131.(84) Coridan, R. H.; Shaner, M.; Wiggenhorn, C.; Brunschwig, B. S.;Lewis, N. S. Electrical and Photoelectrochemical Properties of WO3/SiTandem Photoelectrodes. J. Phys. Chem. C 2013, 117, 6949−6957.(85) Hardee, K. L.; Bard, A. J. Semiconductor Electrodes: X .Photoelectrochemical Behavior of Several Polycrystalline Metal OxideElectrodes in Aqueous Solutions. J. Electrochem. Soc. 1977, 124, 215−224.(86) Koffyberg, F. P.; Benko, F. A. A PhotoelectrochemicalDetermination of the Position of the Conduction and Valence BandEdges of p-Type CuO. J. Appl. Phys. 1982, 53, 1173−1177.(87) Leygraf, C.; Hendewerk, M.; Somorjai, G. A. PhotocatalyticProduction of Hydrogen from Water by a p- and n-TypePolycrystalline Iron Oxide Assembly. J. Phys. Chem. 1982, 86, 4484−4485.(88) Matsumoto, Y.; Sugiyama, K.; Sato, E. Photocathodic HydrogenEvolution Reactions at p-Type CaFe2O4 Electrodes with Fermi LevelPinning. J. Electrochem. Soc. 1988, 135, 98−104.(89) Ida, S.; Yamada, K.; Matsunaga, T.; Hagiwara, H.; Matsumoto,Y.; Ishihara, T. Preparation of p-Type CaFe2O4 Photocathodes forProducing Hydrogen from Water. J. Am. Chem. Soc. 2010, 132,17343−17345.(90) Mor, G. K.; Varghese, O. K.; Wilke, R. H. T.; Sharma, S.;Shankar, K.; Latempa, T. J.; Choi, K.-S.; Grimes, C. A. p-Type Cu−Ti−O Nanotube Arrays and Their Use in Self-Biased HeterojunctionPhotoelectrochemical Diodes for Hydrogen Generation. Nano Lett.2008, 8, 1906−1911.(91) Siripala, W.; Ivanovskaya, A.; Jaramillo, T. F.; Baeck, S.-H.;McFarland, E. W. A Cu2O/TiO2 Heterojunction Thin Film Cathodefor Photoelectrocatalysis. Sol. Energy Mater. Sol. Cells 2003, 77, 229−237.(92) Paracchino, A.; Laporte, V.; Sivula, K.; Gratzel, M.; Thimsen, E.Highly Active Oxide Photocathode for Photoelectrochemical WaterReduction. Nat. Mater. 2011, 10, 456−461.(93) Paracchino, A.; Mathews, N.; Hisatomi, T.; Stefik, M.; Tilley, S.D.; Gratzel, M. Ultrathin Films on Copper(I) Oxide Water SplittingPhotocathodes: A Study on Performance and Stability. Energy Environ.Sci. 2012, 5, 8673−8681.(94) Lin, C.-Y.; Lai, Y.-H.; Mersch, D.; Reisner, E. Cu2O|NiOx

Nanocomposite as an Inexpensive Photocathode In Photoelectro-chemical Water Splitting. Chem. Sci. 2012, 3, 3482−3487.(95) Zhang, Z.; Dua, R.; Zhang, L.; Zhu, H.; Zhang, H.; Wang, P.Carbon-Layer-Protected Cuprous Oxide Nanowire Arrays for EfficientWater Reduction. ACS Nano 2013, 7, 1709−1717.(96) Chen, Z. B.; Jaramillo, T. F.; Deutsch, T. G.; Kleiman-Shwarsctein, A.; Forman, A. J.; Gaillard, N.; Garland, R.; Takanabe, K.;Heske, C.; Sunkara, M.; McFarland, E. W.; Domen, K.; Miller, E. L.;Turner, J. A.; Dinh, H. N. Accelerating Materials Development forPhotoelectrochemical Hydrogen Production: Standards for Methods,Definitions, and Reporting Protocols. J. Mater. Res. 2010, 25, 3−16.(97) Neumann, B.; Bogdanoff, P.; Tributsch, H. TiO2-ProtectedPhotoelectrochemical Tandem Cu(In,Ga)Se2 Thin Film Membranefor Light-Induced Water Splitting and Hydrogen Evolution. J. Phys.Chem. C 2009, 113, 20980−20989.(98) Stavrides, A. Use of Amorphous Silicon Tandem Junction SolarCells for Hydrogen Production in a Photoelectrochemical Cell. Proc.SPIE 2006, 6340, 63400K.(99) Miller, E.; Rocheleau, R. E.; Khan, S. U. M. A HybridMultijunction Photoelectrode for Hydrogen Production Fabricatedwith Amorphous Silicon/Germanium and Iron Oxide Thin Films. Int.J. Hydrogen Energy 2004, 29, 907−914.



(100) van de Krol, R.; Liang, Y. An n-Si/n-Fe2O3 HeterojunctionTandem Photoanode for Solar Water Splitting. Chimia 2013, 67, 168−171.(101) O’Regan, B.; Gratzel, M. A Low-Cost, High-Efficiency SolarCell Based on Dye-Sensitized Colloidal TiO2 Films. Nature 1991, 353,737−740.(102) Gratzel, M. Photoelectrochemical Cells. Nature 2001, 414,338−344.(103) Salvador, P.; Hidalgo, M. G.; Zaban, A.; Bisquert, J.Illumination Intensity Dependence of the Photovoltage in Nano-structured TiO2 Dye-Sensitized Solar Cells. J. Phys. Chem. B 2005, 109,15915−15926.(104) Augustynski, J.; Calzaferri, G.; Courvoisier, J. C.; Gratzel, M. InHydrogen Energy Progress XI: Proceedings of the 11th World HydrogenEnergy Conference, Stuttgart, Germany, 1996, 3, 2379−2387.(105) Gratzel, M. The Artificial Leaf, Bio-Mimetic Photocatalysis.CATTECH 1999, 3, 3−17.(106) Park, J. H.; Bard, A. J. Photoelectrochemical Tandem Cell withBipolar Dye-Sensitized Electrodes for Vectorial Electron Transfer forWater Splitting. Electrochem. Solid-State Lett. 2006, 9, E5−E8.(107) Arakawa, H.; Shiraishi, C.; Tatemoto, M.; Kishida, H.; Usui,D.; Suma, A.; Takamisawa, A.; Yamaguchi, T. Solar HydrogenProduction by Tandem Cell System Composed of Metal OxideSemiconductor Film Photoelectrode and Dye-Sensitized Solar Cell.Solar Hydrogen and Nanotechnology II 2007, 6650, 65003−65003.(108) Nazeeruddin, M. K.; Pechy, P.; Renouard, T.; Zakeeruddin, S.M.; Humphry-Baker, R.; Cointe, P.; Liska, P.; Cevey, L.; Costa, E.;Shklover, V.; Spiccia, L.; Deacon, G. B.; Bignozzi, C. A.; Gratzel, M.Engineering of Efficient Panchromatic Sensitizers for NanocrystallineTiO2-Based Solar Cells. J. Am. Chem. Soc. 2001, 123, 1613−1624.(109) Yum, J. H.; Walter, P.; Huber, S.; Rentsch, D.; Geiger, T.;Nuesch, F.; De Angelis, F.; Gratzel, M.; Nazeeruddin, M. K. EfficientFar Red Sensitization of Nanocrystalline TiO2 Films by anUnsymmetrical Squaraine Dye. J. Am. Chem. Soc. 2007, 129, 10320−10321.(110) Shi, Y.; Hill, R. B. M.; Yum, J.-H.; Dualeh, A.; Barlow, S.;Gratzel, M.; Marder, S. R.; Nazeeruddin, M. K. A High-EfficiencyPanchromatic Squaraine Sensitizer for Dye-Sensitized Solar Cells.Angew. Chem. 2011, 123, 6749−6751.(111) Brillet, J.; Cornuz, M.; Le Formal, F.; Yum, J. H.; Graetzel, M.;Sivula, K. Examining Architectures of Photoanode-PhotovoltaicTandem Cells for Solar Water Splitting. J. Mater. Res. 2010, 25, 17−24.(112) Yum, J.-H.; Baranoff, E.; Kessler, F.; Moehl, T.; Ahmad, S.;Bessho, T.; Marchioro, A.; Ghadiri, E.; Moser, J.-E.; Yi, C.;Nazeeruddin, M. K.; Gratzel, M. A Cobalt Complex Redox Shuttlefor Dye-Sensitized Solar Cells with High Open-Circuit Potentials. Nat.Commun. 2012, 3, 631.(113) Brillet, J.; Yum, J. H.; Cornuz, M.; Hisatomi, T.; Solarska, R.;Augustynski, J.; Graetzel, M.; Sivula, K. Highly Efficient WaterSplitting by a Dual-Absorber Tandem Cell. Nat. Photonics 2012, 6,824−828.(114) Delcamp, J. H.; Yella, A.; Holcombe, T. W.; Nazeeruddin, M.K.; Gratzel, M. The Molecular Engineering of Organic Sensitizers forSolar-Cell Applications. Angew. Chem., Int. Ed. 2013, 52, 376−380.



Documents

Photoelectrochemical Tandem Cells for Solar Water Splittingcourses.washington.edu/mengr599/notes/Journal-Article-3-2-2016.pdf · potentials against the reversible hydrogen electrode,